Muzzle Blast Detection via Short Time Fourier Transform

A projectile's trajectory can be estimated by measuring the arrival times of the acoustic energy at several locations in space. In case of a supersonic projectile fired from a gun, both the acoustic shockwave and muzzle blast can be observed. For acoustic sensor networks attempting to determine the trajectory, the challenge is to first, correctly classify the transient signal as either a shockwave or a muzzle blast and then, calculate the direction of arrival based on arrival times across a sensor array. This can be challenging when the shockwave has lost substantial high frequency content. Various joint time-frequency techniques can be used to classify these signatures, such as short-time Fourier transform (STFT), Smoothed Pseudo Wigner-Ville distribution (SPWVD) and wavelet analysis [1]. This paper presents a short-time Fourier transform (STFT) technique that was implemented in a shooter localization application. The application is built on a sensor network developed by the Network Embedded Systems Technology group at Institute for Software Integrated Systems, at the Vanderbilt University [2]. The ad-hoc sensor network consists of a large number of independent motes, each equipped with a multi-purpose acoustic sensor board (MicroBee). The MicroBee board has two independent acoustic channels (each with a microphone, amplifier with controllable gain and ADC operating up to 100 kHz) and an ADSP-2189 16 bit fixed-point DSP processor. The presented technique focuses on the muzzle blast detection, however, similar techniques can be used also for the shockwave signature detection. Elements of the algorithms proposed by [3] are utilized in the presented technique with more emphasis laid on realization aspects.